Cue relevance and feedback in a clinical prediction task

Abstract

An experiment was performed on the effect on individual judges' predictive efficiency of information about the relevance of cues for clinical judgments. The information patterns were such that the task may be characterized as a multiple-cue probability task. The predictive efficiency was shown to be substantially and significantly higher with access to the ecological validity coefficients than without such information. Feedback in the form of utilization coefficients and functional validities had no additional positive effect.